topic nested/more efficient macro for proc logistic model in Statistical Procedures

nested/more efficient macro for proc logistic

Caetreviop543 — Tue, 29 Aug 2017 04:16:51 GMT

I have two different data sets, six different outcome variables, and two primary (separate) predictors which I am trying to incorporate into proc logistic along with several covariates using a macro. I'm trying to figure out the most efficient way. My initial thought was:

%macro mod(dat, out, pred);

proc logistic data=&dat; 
model &out=&pred cov1 cov2 cov3 cov4 cov5 cov6 /desc;
run;

%mend mod;
%mod (dat1, out1, pred1);
%mod (dat1, out2, pred1);
%mod (dat1, out3, pred1);
%mod (dat1, out4, pred1);
%mod (dat1, out5, pred1);
%mod (dat1, out6, pred1);
%mod (dat1, out1, pred2);
%mod (dat1, out2, pred2);
%mod (dat1, out3, pred2);
%mod (dat1, out4, pred2);
%mod (dat1, out5, pred2);
%mod (dat1, out6, pred2);
%mod (dat2, out1, pred1);
%mod (dat2, out2, pred1);
%mod (dat2, out3, pred1);
%mod (dat2, out4, pred1);
%mod (dat2, out5, pred1);
%mod (dat2, out6, pred1);
%mod (dat2, out1, pred2);
%mod (dat2, out2, pred2);
%mod (dat2, out3, pred2);
%mod (dat2, out4, pred2);
%mod (dat2, out5, pred2);
%mod (dat2, out6, pred2);

Is there a more efficient way, by referring to the outcome variables only once/nesting the macro? Another, however also inefficient, way is:

%macro mod(out);

proc logistic data=data_one;
model &out=pred1 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

proc logistic data=data_one;
model &out=pred2 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

proc logistic data=data_two;
model &out=pred1 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

proc logistic data=data_two;
model &out=pred2 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

%mend mod;
%mod (out1);
%mod (out2);
%mod (out2);
%mod (out3);
%mod (out4);
%mod (out5);
%mod (out6);

Thanks in advance!

nested/more efficient macro for proc logistic model

Caetreviop543 — Tue, 29 Aug 2017 04:19:39 GMT

favorite

Hello,

I have two data sets, six different outcome variables, and two primary (separate) predictors which I am trying to incorporate into proc logistic along with several covariates using a macro. I'm trying to figure out the most efficient way. My initial thought was:

%macro mod(dat, out, pred);

proc logistic data=&dat; 
model &out=&pred cov1 cov2 cov3 cov4 cov5 cov6 /desc;
run;

%mend mod;
%mod (dat1, out1, pred1);
%mod (dat1, out2, pred1);
%mod (dat1, out3, pred1);
%mod (dat1, out4, pred1);
%mod (dat1, out5, pred1);
%mod (dat1, out6, pred1);
%mod (dat1, out1, pred2);
%mod (dat1, out2, pred2);
%mod (dat1, out3, pred2);
%mod (dat1, out4, pred2);
%mod (dat1, out5, pred2);
%mod (dat1, out6, pred2);
%mod (dat2, out1, pred1);
%mod (dat2, out2, pred1);
%mod (dat2, out3, pred1);
%mod (dat2, out4, pred1);
%mod (dat2, out5, pred1);
%mod (dat2, out6, pred1);
%mod (dat2, out1, pred2);
%mod (dat2, out2, pred2);
%mod (dat2, out3, pred2);
%mod (dat2, out4, pred2);
%mod (dat2, out5, pred2);
%mod (dat2, out6, pred2);

Is there a more efficient way, by referring to the outcome variables only once/nesting the macro? Another, however also inefficient, way is:

%macro mod(out);

proc logistic data=data_one;
model &out=pred1 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

proc logistic data=data_one;
model &out=pred2 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

proc logistic data=data_two;
model &out=pred1 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

proc logistic data=data_two;
model &out=pred2 cov1 cov2 cov3 cov4 cov5 cov6/desc;
run;

%mend mod;
%mod (out1);
%mod (out2);
%mod (out2);
%mod (out3);
%mod (out4);
%mod (out5);
%mod (out6);

Thanks in advance!

Re: nested/more efficient macro for proc logistic

Reeza — Tue, 29 Aug 2017 04:47:13 GMT

Or, don't write a macro and transform your data and use a BY statement.

https://communities.sas.com/t5/SAS-Communities-Library/How-do-I-write-a-macro-to-run-multiple-regressions/ta-p/223663

Re: nested/more efficient macro for proc logistic

Caetreviop543 — Tue, 29 Aug 2017 04:58:33 GMT

That's interesting! I wouldn't have thought to do it that way...

Thanks Reeza!